The Research Of Highway Vertical Alignment Optimization Using Genetic Algorithms

After the highway horizontal alignment is fixed, the vertical alignment determination is the sensitive element of highway design. The vertical alignment has important implications not only on road construction costs but also with respect to disruption of natural landform and soil conservation. Optimizing the vertical alignment is very necessary in highway design. So, an evolutionary model using the improved genetic algorithms (GAs) has been developed in order to seek the most suitable solution for the earthwork cost objective. The model tries to increase the quality and efficiency of the vertical alignment design in the practice.In order to optimize the stake and elevation of grade change points simultaneously, this model chose the combination of grade change points as the decision variable during the optimizing process. With the calculated earthwork volume as the model objective, the fill-cut cost coefficient was created in the model to express the fill or cut trend in the vertical alignment design.Optimal vertical alignment analysis was made and some key problems during the GAs application were resolved primarily. According to the diversity evaluation of the population, the happening probability of selection operator, crossover operator and mutation operator was self-adjusted respectively during the population evolution and the mutation operator was non-uniform. The original grade change points were generated randomly by the normal distribution whose expectation was the corresponding average-stake value. As far as the constraint violation during optimal searching, the constraint processing strategy and method were also discussed in this paper.After the detailed design of the genetic algorithms, a highway vertical alignment optimal program using the improved genetic algorithms was achieved. And the program proved available to some extend by the optimal results of an example in different cases.